A novel bispecific, trivalent antibody construct for targeting pancreatic carcinoma.

Preclinical and clinical studies have demonstrated the application of radiolabeled mAb-PAM4 for nuclear imaging and radioimmunotherapy of pancreatic carcinoma. We have now examined the ability of a novel PAM4-based, bispecific monoclonal antibody (mAb) construct, TF10, to pretarget a radiolabeled peptide for improved imaging and therapy. TF10 is a humanized, bispecific mAb, divalent for mAb-PAM4 and monovalent for mAb-679, reactive against the histamine-succinyl-glycine hapten. Biodistribution studies and nuclear imaging of the radiolabeled TF10 and/or TF10-pretargeted hapten-peptide (IMP-288) were conducted in nude mice bearing CaPan1 human pancreatic cancer xenografts. (125)I-TF10 cleared rapidly from the blood, with levels decreasing to <1% injected dose per gram (ID/g) by 16 hours. Tumor uptake was 3.47 +/- 0.66% ID/g at this time point with no accumulation in any normal tissue. To show the utility of the pretargeting approach, (111)In-IMP-288 was administered 16 hours after TF10. At 3 hours postadministration of radiolabeled peptide, imaging showed intense uptake within the tumors and no evidence of accretion in any normal tissue. No targeting was observed in animals given only the (111)In-peptide. Tumor uptake of the TF10-pretargeted (111)In-IMP-288 was 24.3 +/- 1.7% ID/g, whereas for (111)In-IMP-288 alone it was only 0.12 +/- 0.002% ID/g at 16 hours. Tumor/blood ratios were significantly greater for the pretargeting group ( approximately 1,000:1 at 3 hours) compared with (111)In-PAM4-IgG ( approximately 5:1 at 24 hours; P < 0.0003). Radiation dose estimates suggested that TF10/(90)Y-peptide pretargeting would provide a greater antitumor effect than (90)Y-PAM4-IgG. Thus, the results suggest that TF10 pretargeting may provide improved imaging for early detection, diagnosis, and treatment of pancreatic cancer as compared with directly radiolabeled PAM4-IgG.